spacer Persistence of Vision

Persistence of Vision (POV) displays are fun to watch but can be challenging to build. The common explanations for why they work are apparently based on a myth. Luckily they DO still work. If you sweep a light source through the air quickly enough you will fool the brain into thinking that the light exists in multiple places at once.

Typical POV displays take a linear array of LEDs and sweep them through the air with a circular motion to produce either a disk or a cylindrical region. By turning the LEDs on and off during the motion, we can produce a 2D display using a simple 1D array of LEDs.

I was drawn to an old Fujitsu hard drive that I had lying about. The drive had seen better days but I knew that within its case lay a near frictionless, robust spindle that ought to work wonderfully for my fledgling POV project.

At this point I had not yet been introduced to synchronous DC motors. If you have played with these you may snicker now in anticipation of my woes.

Fujitsu HD without cover

The motor on this drive has four wires. The coils inside are arranged in a Y configuration. In my case I connected the common lead to a voltage source (~9 - 12V) and alternately grounded the various leads A, B and C. With current flowing from the common lead into any one of the other leads, the rotor swings to one of four positions spaced 90 degrees apart and slowly wobbles to a stop over a few seconds. The picture below shows 'my imagination' of what is happening (having not actually looked inside). The other two leads correspond to similar positions that are shifted 30 degrees from each other.

Fujitsu Hard Driver Motor - wireside viewY Shaped arrangement of coils in motor

So, to make this motor turn, you need to energize lead A, then B, and then C in sequence (then repeat). If the timing is perfect, the motor with rotate smoothly. For my first naive attempt I used my PIC to produce a slowly accelerating sequence of pulses that did exactly this. And did it work? Well, sort of (not). At times the motor would accelerate haltingly and synchronize for a while only to stop without provocation, while other times it would just jitter about.

Resonance is the key. Everyone knows that to push a kid on a swing you need to push at exactly the right moment. Too soon or too late and soon you have a complaining kid on your hands because they're not swinging higher. The synchronous DC motor is the same. If you can give it a perfectly timed nudge, it will accelerate smoothly and stubbornly hold whatever speed you want. Miss your timing by a smidge and soon it jitters to a stop. And this calls for feedback.

How does the hard drive do it? My suspicion is that the magnetic information on the disk itself is used to provide this feedback. There are no other obvious sources to be seen in the disassembled drive. Further, the driver chips I've seen combine motor and read/write head control so it would make a lot of sense.

A colleague of mine claims you can use back emf to provide this feedback. No doubt he is correct, and a cleverer person than I could make it work. In my case, I resorted to an optical encoder because it reeks of digital, and digital is what I love. My optical encoder was drawn in Corel Draw and printed on a laser transparency. Not hi-tech but, in conjunction with a small optical interrupt switch, it worked really very well.

Next Section: Synchronous DC Motor Driver